OpenAI launching election war room
The developer will monitor abuse of its AI tools during next week's midterm elections, as bots and deepfakes proliferate.
A group of select OpenAI employees will spend Nov. 3 in the AI giant's new midterm election "war room," tasked with monitoring ChatGPT's role in the political machine.
The company announced, via Politico, that it would convene its own Election Operations Center ahead of next week's nationwide elections with the express task of spotting potential misuse of its AI models.
SEE ALSO: NY governor signs major chatbot safety bill into law"We’ve put considerable effort [into] ensuring we help users get reliable voting information, combat misuse of our tools, and support election defenders. The Elections Operations Center builds on that years long work, bringing our teams together to respond quickly to emerging issues and enforce our policies as people head to the polls," Josh Lawson, head of civic and strategic risk for OpenAI, told Decoded by Politico.
Other tech giants have diverted resources to election integrity in recent years, including Microsoft's voter education and AI literacy initiative and Meta's $30 billion spend to advance its election security initiatives, following allegations its platforms sowed political discord.
In the midst of the 2024 presidential election, a coalition of tech companies pledged to work together to combat generative AI's potential to undermine global elections. The agreement included commitments to mass public education, more transparent AI policies, and robust detection tools to curb the spread of political deepfakes.
Even so, many remain concerned about AI's role in election disinformation campaigns. According to misinformation watchdog NewsGuard, disinformation campaigns directed by foreign governments have increased ahead of election day, primarily amplified on social media platforms like X. More than half of Americans were fooled by such election misinformation, and X's Community Notes failed to identify more than 93 percent of such posts, according to the organization's research.
A report from the Brennan Center for Justice found that guardrails in major AI models, including Gemini and Meta AI, were failing to prevent the generation of fake election images — ChatGPT was the only model that showed some improvement in curbing photorealistic images, according to the center's findings. But NewsGuard's tests of OpenAI's ChatGPT Images 2.5 found the model repeatedly generated lifelike images of political candidates when prompted.
Meanwhile, a growing number of Americans are leaning on chatbots to help them find polling information, research candidates, and even fill out their ballots.
SpaceXSI (formerly xAI) owner Elon Musk has personally contributed more than $366 million to Republican election efforts ahead of the 2026 election, making him one of the most prolific mega donors in recent history.